euge.tv

After a long break I’ve begun blogging at http://eugeneware.com

You can add a bookmark to your toolbar that will allow you to blog post the current window. You need to remove all the line breaks to put this code on one line. Also, you need to change www.digitalvideosecrets.com to the url of your drupal installation

javascript:u=document.location.href; t=document.title.replace(/\|/g,'::');
s='<blockquote>'+window.getSelection()+'</blockquote>';
pre='<em>From <a target="_blank" href="'+escape(u)+'">'+escape(t) + 
escape('</a>:</em><br /> '); r />
w = (window.open('http://www.digitalvideosecrets.com/node/add/blog?edit[title]='
+escape('Link: '+t)+'&edit[body]='+pre+escape(s)+
escape('<a href="'+u+'" target="_blank">Link</a>'),'_blank','width=710,height=500,
status=yes,resizable=yes,scrollbars=yes'));
void(window.setTimeout(function(){w.focus()},300));

I’ve added a meebo chat box, so if you see me online, then you’ll be able to chat to me in real time!!!

Just enter in the euge.tv box at the right of this page, and if I’m online I’ll answer…

So cool!

If you haven’t read my previous article, you may wish to read part one. In it I defined the 8 key reasons why video looks different to film.

I also put together an example of some raw video, that looks like video, and then I used some tricks (which I’ll show you in due course) to make the video footage look more “film like”.

If you haven’t already, I highly recommend that you watch the video showing the comparison. The difference is quite startling, particular once you consider that you’re looking at exactly the same footage.

The importance of frame-rate in producing “Film Look” video

In this article I’m going to go over the first “Film Look” secret. And that’s framerate.

As I touched on in my first article, frame rate is the number of images that get displayed per second.

For NTSC video this is 29.97 frames per second.

For PAL video this is 25.00 frames per second.

But for film, the correct figure is 23.976 frames per second.

The slower frame rate of film also creates more “motion blur”, so that each frame is not so “sharp” while in motion.

Whether consciously, or otherwise, this is a lot of the reason why when you look at film (or even video that originated from film) that you associate it with cinema.

Interlaced or Progressive – that is the question…

The other significant difference between film and video is that video is interlaced (showing alternate even and odd scan lines on the television), while film is progressive (full frames are captured and displayed when projected in the cinema).

As a result, PAL 25 frames per second video is actually 50 fields per second that gets broadcast from the stations, and captured to video tape. For that reason, PAL video is also typically referred to as 50i (50 interlaced fields per second) video.

And similar NTSC 29.97 frames per second is actually 59.94 fields per second. And because people can’t be bothered spelling out 59.94 they usually refer to NTSC video as 60i (60 interlaced fields per second) video.

And while we’re on the subject of framerates and fieldrates, the 23.976 progressive framerate of video is usually abbreviated to 24p (23.976 progressive frames per second) video.

Still with me? Suffice to say that the interlaced nature of video with a different framerate is a big part of the “video look” that many people hate (particularly if you’re an indie film maker!)

How do you create 24 frames-per-second video footage?

So, how do we change our 25fps (frames per second) and 29.97fps footage to be 23.976fps?

Good question. And the answer is not straightforward. You have a couple of options:

1. Shoot at 24p using a camera that natively supports this framerate.
2. Shoot at 50i, convert to 25p (25 progressive frames per second), then slow down the footage (and audio as well) by 4% to get it to 24p
3. Shoot at 60i, convert to 60p, and blend, interpolate, combine frames to get a motion-blurred 24p.

Option 1: Shoot 24p

The first option is the easiest, and most preferable way to create the “film look”, however, you need to get a camera that natively records in 24p. One popular camera that does this is the Panasonic DVX100. Because the whole camera is set up to record full progressive frames on it’s video sensors, a very high quality progressive picture is recorded and it produces the highest quality result.

However, cameras that record 24p are not available in the consumer camcorder category, and are at least $3,000 to purchase. 24p cameras also usually support “film like” color reproduction and light response (gamma), and are really the best choice if you are serious about producing a quality “film look” feature.

Option 2: Shoot PAL and convert

The second option is probably the best trade-off. You take 50i footage, “de-interlace” the footage to create 25p footage. If your target audience is PAL, then just use the 25p footage, as 25p is virtually indistinguishable to 24p to the end-user.

In fact, that’s how film is transferred to television and DVD in most cases in PAL countries. Studios scan the film stock at 24p, then speed it up by 4% to get to 25p, which is then duplicated to create 50i footage that is suitable for TV or DVD. The pitch of the audio is increased by 4% and is therefore slightly higher than the original, but most people would not perceive the difference unless they had the original audio to compare with.

If you are wanting to target an NTSC audience then you will have to go from 25p to 24p, slowing down the footage by 4% and lowering the pitch of the audio. Then you need to apply a process called “3:2 pulldown” to convert the 24p footage to 60i, suitable for NTSC televisions.

This sounds complicated, and in one sense it is. Basically you have to translate 24 frames per second to 30 frames per second (ie. 60 fields per second). This is achieved by turning every 2 frames from the 24p footage into 5 fields in the 60i footage. Therefore 24 frames (twelve times this) would result in twelve times five, or the full 60 fields present in the NTSC 60i signal.

Converting two 24p frames into five 60i fields means duplicating the first 24p frame into two 60i fields, and the second 24p frame is actually duplicated into the next three 60i fields. This process is known as adding a 3:2 pulldown. Because the way that the fields get “pulled down” is uneven, this is referred to as a 3:2 pulldown “judder”, as there is one too many fields in every five fields of video.

So, in summary, once you have your 24p footage from PAL video, you convert it to NTSC by adding the 3:2 pulldown to convert it to 60i video. In addition, because PAL has a higher vertical resolution (720×576) than NTSC (720×480), the vertical size of the video is also adjusted.

Option 3: Shoot NTSC and convert

The third option, which is to shoot in 60i NTSC and convert to 24p is probably the most complicated move. It certainly is possible, but of all the three methods, it is usually the most involved. There are many methods that can be applied varying from reversing 3:2 pulldown to recreate progressive frames, to converting to 60p (or even 120p) footage and then blending, interpolating and combining frames using sophisticated algorithms to get to an approximation of 24p footage, and also the characteristic 24p motion blur.

So, practically, how do you create 24p?

Ok, so that’s the theory of how to create 24p. But how in practice do you actually do it?

Well, if you’re after purchasing a 24p camera I’d recommend purchasing a Panasonic DVX100 for standard-definition content, or the much more expensive HVX200 for high-definition video.

If you’re reading this article, then chances are that you already own a video camera. In this way, regardless of whether you own an NTSC camera, or a PAL camera, I’d highly recommend purchasing a copy of DVFilm Maker

DVFilm Maker has all the smarts to convert your footage to 24P regardless of what time of video format it is. And for $145 it’s quite inexpensive.

Or, if you’re on a real budget and you’re prepared to put in a bit of time to learn some new tools and even write some “code”, I’d highly recommend learning AVISynth

AviSynth is a scripting language that enables you to manipulate video at a very low level.

When combined with the fabulous open source tool virtualdub and some user contributed scripts and the help of the Doom9 Community you will be able to do anything you can imagine with your footage.

However, AviSynth is not for the faint-hearted, and for the majority of people I’d recommend using an off-the-shelf tool such as DVFilm Maker.

For the more serious film-maker with a larger budget you might like to take a look at Magic Bullet

Conclusion

The entire independent film-making community changed almost overnight with the release of the first 24p camera, that enabled the shooting of “film look” footage on digital video.

The “film look” attributable to the 24p “cadence” of film can now be reproduced somewhat by either shooting in 24p or converting existing footage to 24p using software such as DVFilm Maker and Magic Bullet.

There’s much more to making video seem film-like than just 24p, and we’ll cover more of these techniques in upcoming articles.

Did you ever wonder why video looks like video, and film looks like film? Or why the 6 o’clock news doesn’t look like a movie?

Or, put another way, what can you do to make your video appear more cinematic?

Here’s the key differences:

1. Framerate – Film is shot at a framerate of 24 (actually 23.976) frames a second, video is either 25 or 30 (actually 29.97)
2. Progressive vs Interlaced – Film is progressive, in that 23.976 full frames are recorded to film each second. With NTSC video, for example, there are actually 59.94 (ie. 2×29.97) fields that get displayed a second. Interlacing means that only half the picture is shown alternatively. So first the even lines are displayed, then the odd-lines, then the even lines, then the odd, and so on. This “interlaced” appearance is often a big tip-off that you have shot on video.
3. Depth-of-field – Most cinematic productions make use of a shallow “depth-of-field” to focus the audience’s attention on objects and people. This has the appearance of keeping something in focus while the background or foreground is blurred out, leading to a pleasing image. Video is typically shot on lenses that provide a very wide depth-of-field where everything is in focus.
4. Exposure Latitute – Film has a greater ability to represent both very dark images and very bright images in the same shot due to it’s very high exposure latitude. Video, however, is not so flexible. If you have both very dark and light subjects that you wish to capture with video you’ll need to either expose for the dark subject and all the bright areas will be “blown out” and just appear too bright and washed out, or you expose for the bright item, and you lose all the detail in the dark areas.
5. Production Value – Most cinematic features have much higher budgets than television or other video formats. This manifests itself in smooth camera and crane moves, and also much better or artistic lighting than is typically found on video.
6. Aspect Ratio – The majority of television and video is still watched on a 4×3 aspect ratio. Even though more and more television is being released in 16×9 widescreen, most film is recorded these days in 2:35:1 Panavision which is wider again.
7. Gamma – Because Film is a very difference medium to video, it responds very differently to light. This often referred to as the “gamma” of the medium. Film is referred to as having a “cinegamma”.
8. Color – In general you’ll find that colors on film look richer and more saturated than on video.

I you want to achieve this coveted “film look” there are many things you can do. I’ll cover some of the easy “post production” ways in the next part in this series.

In the meantime, you can checkout a comparison between pure untreated video and video that has been tweaked to look more like film by clicking here